From 27a7d8d5f711dad8bddc67eceb0387a00c7e11e8 Mon Sep 17 00:00:00 2001
From: Kiva <imkiva@islovely.icu>
Date: Mon, 25 Mar 2024 08:02:46 +0800
Subject: [PATCH] [Clang][XTHeadVector] implement 12.13
 vwmaccu/vwmacc/vwmaccsu/vwmaccus (#87)

* [Clang][XTHeadVector] implement 12.13 `vwmaccu/vwmacc/vwmaccsu/vwmaccus`

* [Clang][XTHeadVector] test 12.13 `vwmaccu/vwmacc/vwmaccsu/vwmaccus`
---
 .../clang/Basic/riscv_vector_xtheadv.td       |  20 +
 .../vector-widening-mul-add/thead/vwmacc.c    | 367 ++++++++++++++++++
 .../vector-widening-mul-add/thead/vwmaccsu.c  | 367 ++++++++++++++++++
 .../vector-widening-mul-add/thead/vwmaccu.c   | 367 ++++++++++++++++++
 .../vector-widening-mul-add/thead/vwmaccus.c  | 187 +++++++++
 5 files changed, 1308 insertions(+)
 create mode 100644 clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmacc.c
 create mode 100644 clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccsu.c
 create mode 100644 clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccu.c
 create mode 100644 clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccus.c

diff --git a/clang/include/clang/Basic/riscv_vector_xtheadv.td b/clang/include/clang/Basic/riscv_vector_xtheadv.td
index f89b31ee28da1b..a85f6da4475202 100644
--- a/clang/include/clang/Basic/riscv_vector_xtheadv.td
+++ b/clang/include/clang/Basic/riscv_vector_xtheadv.td
@@ -61,6 +61,10 @@ multiclass RVVOp0Op1BuiltinSet<string intrinsic_name, string type_range,
                                list<list<string>> suffixes_prototypes>
     : RVVBuiltinSet<intrinsic_name, type_range, suffixes_prototypes, [0, 1]>;
 
+multiclass RVVOutOp1Op2BuiltinSet<string intrinsic_name, string type_range,
+                                  list<list<string>> suffixes_prototypes>
+    : RVVBuiltinSet<intrinsic_name, type_range, suffixes_prototypes, [-1, 1, 2]>;
+
 multiclass RVVSignedBinBuiltinSet
     : RVVOutOp1BuiltinSet<NAME, "csil",
                           [["vv", "v", "vvv"],
@@ -1122,6 +1126,22 @@ let MaskedPolicyScheme = NonePolicy,
 }
 
 // 12.13. Vector Widening Integer Multiply-Add Operations
+let MaskedPolicyScheme = NonePolicy,
+    UnMaskedPolicyScheme = NonePolicy,
+    HasMaskedOffOperand = false,
+    Log2LMUL = [-3, -2, -1, 0, 1, 2] in {
+  defm th_vwmaccu : RVVOutOp1Op2BuiltinSet<"th_vwmaccu", "csi",
+                                           [["vv", "Uw", "UwUwUvUv"],
+                                            ["vx", "Uw", "UwUwUeUv"]]>;
+  defm th_vwmacc : RVVOutOp1Op2BuiltinSet<"th_vwmacc", "csi",
+                                          [["vv", "w", "wwvv"],
+                                           ["vx", "w", "wwev"]]>;
+  defm th_vwmaccsu : RVVOutOp1Op2BuiltinSet<"th_vwmaccsu", "csi",
+                                            [["vv", "w", "wwvUv"],
+                                             ["vx", "w", "wweUv"]]>;
+  defm th_vwmaccus : RVVOutOp1Op2BuiltinSet<"th_vwmaccus", "csi",
+                                            [["vx", "w", "wwUev"]]>;
+}
 
 // 12.14. Vector Integer Merge Operations
 
diff --git a/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmacc.c b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmacc.c
new file mode 100644
index 00000000000000..18dc89ede393cb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmacc.c
@@ -0,0 +1,367 @@
+// RUN: %clang_cc1 -triple riscv64 -target-feature +xtheadvector \
+// RUN:   -disable-O0-optnone -emit-llvm %s -o - | \
+// RUN:   opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmacc_vv_i16m2
+// CHECK-RV64-SAME: (<vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x i8> [[VS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmacc.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x i8> [[VS1]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmacc_vv_i16m2(vint16m2_t vd, vint8m1_t vs1, vint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmacc_vx_i16m2
+// CHECK-RV64-SAME: (<vscale x 8 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmacc.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmacc_vx_i16m2(vint16m2_t vd, int8_t rs1, vint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmacc_vv_i16m4
+// CHECK-RV64-SAME: (<vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x i8> [[VS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmacc.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x i8> [[VS1]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmacc_vv_i16m4(vint16m4_t vd, vint8m2_t vs1, vint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmacc_vx_i16m4
+// CHECK-RV64-SAME: (<vscale x 16 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmacc.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmacc_vx_i16m4(vint16m4_t vd, int8_t rs1, vint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmacc_vv_i16m8
+// CHECK-RV64-SAME: (<vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x i8> [[VS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmacc.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x i8> [[VS1]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmacc_vv_i16m8(vint16m8_t vd, vint8m4_t vs1, vint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmacc_vx_i16m8
+// CHECK-RV64-SAME: (<vscale x 32 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmacc.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmacc_vx_i16m8(vint16m8_t vd, int8_t rs1, vint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmacc_vv_i32m2
+// CHECK-RV64-SAME: (<vscale x 4 x i32> [[VD:%.*]], <vscale x 4 x i16> [[VS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmacc.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], <vscale x 4 x i16> [[VS1]], <vscale x 4 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmacc_vv_i32m2(vint32m2_t vd, vint16m1_t vs1, vint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmacc_vx_i32m2
+// CHECK-RV64-SAME: (<vscale x 4 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmacc.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmacc_vx_i32m2(vint32m2_t vd, int16_t rs1, vint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i32m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmacc_vv_i32m4
+// CHECK-RV64-SAME: (<vscale x 8 x i32> [[VD:%.*]], <vscale x 8 x i16> [[VS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmacc.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], <vscale x 8 x i16> [[VS1]], <vscale x 8 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmacc_vv_i32m4(vint32m4_t vd, vint16m2_t vs1, vint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmacc_vx_i32m4
+// CHECK-RV64-SAME: (<vscale x 8 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmacc.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmacc_vx_i32m4(vint32m4_t vd, int16_t rs1, vint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i32m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmacc_vv_i32m8
+// CHECK-RV64-SAME: (<vscale x 16 x i32> [[VD:%.*]], <vscale x 16 x i16> [[VS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmacc.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], <vscale x 16 x i16> [[VS1]], <vscale x 16 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmacc_vv_i32m8(vint32m8_t vd, vint16m4_t vs1, vint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmacc_vx_i32m8
+// CHECK-RV64-SAME: (<vscale x 16 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmacc.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmacc_vx_i32m8(vint32m8_t vd, int16_t rs1, vint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i32m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmacc_vv_i64m2
+// CHECK-RV64-SAME: (<vscale x 2 x i64> [[VD:%.*]], <vscale x 2 x i32> [[VS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmacc.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], <vscale x 2 x i32> [[VS1]], <vscale x 2 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmacc_vv_i64m2(vint64m2_t vd, vint32m1_t vs1, vint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i64m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmacc_vx_i64m2
+// CHECK-RV64-SAME: (<vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmacc.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmacc_vx_i64m2(vint64m2_t vd, int32_t rs1, vint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i64m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmacc_vv_i64m4
+// CHECK-RV64-SAME: (<vscale x 4 x i64> [[VD:%.*]], <vscale x 4 x i32> [[VS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmacc.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], <vscale x 4 x i32> [[VS1]], <vscale x 4 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmacc_vv_i64m4(vint64m4_t vd, vint32m2_t vs1, vint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i64m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmacc_vx_i64m4
+// CHECK-RV64-SAME: (<vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmacc.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmacc_vx_i64m4(vint64m4_t vd, int32_t rs1, vint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i64m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmacc_vv_i64m8
+// CHECK-RV64-SAME: (<vscale x 8 x i64> [[VD:%.*]], <vscale x 8 x i32> [[VS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmacc.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], <vscale x 8 x i32> [[VS1]], <vscale x 8 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmacc_vv_i64m8(vint64m8_t vd, vint32m4_t vs1, vint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i64m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmacc_vx_i64m8
+// CHECK-RV64-SAME: (<vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmacc.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmacc_vx_i64m8(vint64m8_t vd, int32_t rs1, vint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i64m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmacc_vv_i16m2_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x i8> [[VS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmacc.mask.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x i8> [[VS1]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmacc_vv_i16m2_m(vbool8_t mask, vint16m2_t vd, vint8m1_t vs1, vint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmacc_vx_i16m2_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmacc.mask.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmacc_vx_i16m2_m(vbool8_t mask, vint16m2_t vd, int8_t rs1, vint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmacc_vv_i16m4_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x i8> [[VS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmacc.mask.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x i8> [[VS1]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmacc_vv_i16m4_m(vbool4_t mask, vint16m4_t vd, vint8m2_t vs1, vint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmacc_vx_i16m4_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmacc.mask.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmacc_vx_i16m4_m(vbool4_t mask, vint16m4_t vd, int8_t rs1, vint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmacc_vv_i16m8_m
+// CHECK-RV64-SAME: (<vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x i8> [[VS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmacc.mask.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x i8> [[VS1]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmacc_vv_i16m8_m(vbool2_t mask, vint16m8_t vd, vint8m4_t vs1, vint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmacc_vx_i16m8_m
+// CHECK-RV64-SAME: (<vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmacc.mask.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmacc_vx_i16m8_m(vbool2_t mask, vint16m8_t vd, int8_t rs1, vint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i16m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmacc_vv_i32m2_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i32> [[VD:%.*]], <vscale x 4 x i16> [[VS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmacc.mask.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], <vscale x 4 x i16> [[VS1]], <vscale x 4 x i16> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmacc_vv_i32m2_m(vbool16_t mask, vint32m2_t vd, vint16m1_t vs1, vint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i32m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmacc_vx_i32m2_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmacc.mask.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmacc_vx_i32m2_m(vbool16_t mask, vint32m2_t vd, int16_t rs1, vint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i32m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmacc_vv_i32m4_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i32> [[VD:%.*]], <vscale x 8 x i16> [[VS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmacc.mask.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], <vscale x 8 x i16> [[VS1]], <vscale x 8 x i16> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmacc_vv_i32m4_m(vbool8_t mask, vint32m4_t vd, vint16m2_t vs1, vint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i32m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmacc_vx_i32m4_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmacc.mask.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmacc_vx_i32m4_m(vbool8_t mask, vint32m4_t vd, int16_t rs1, vint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i32m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmacc_vv_i32m8_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i32> [[VD:%.*]], <vscale x 16 x i16> [[VS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmacc.mask.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], <vscale x 16 x i16> [[VS1]], <vscale x 16 x i16> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmacc_vv_i32m8_m(vbool4_t mask, vint32m8_t vd, vint16m4_t vs1, vint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i32m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmacc_vx_i32m8_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmacc.mask.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmacc_vx_i32m8_m(vbool4_t mask, vint32m8_t vd, int16_t rs1, vint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i32m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmacc_vv_i64m2_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i64> [[VD:%.*]], <vscale x 2 x i32> [[VS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmacc.mask.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], <vscale x 2 x i32> [[VS1]], <vscale x 2 x i32> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmacc_vv_i64m2_m(vbool32_t mask, vint64m2_t vd, vint32m1_t vs1, vint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i64m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmacc_vx_i64m2_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmacc.mask.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmacc_vx_i64m2_m(vbool32_t mask, vint64m2_t vd, int32_t rs1, vint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i64m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmacc_vv_i64m4_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i64> [[VD:%.*]], <vscale x 4 x i32> [[VS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmacc.mask.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], <vscale x 4 x i32> [[VS1]], <vscale x 4 x i32> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmacc_vv_i64m4_m(vbool16_t mask, vint64m4_t vd, vint32m2_t vs1, vint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i64m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmacc_vx_i64m4_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmacc.mask.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmacc_vx_i64m4_m(vbool16_t mask, vint64m4_t vd, int32_t rs1, vint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i64m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmacc_vv_i64m8_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i64> [[VD:%.*]], <vscale x 8 x i32> [[VS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmacc.mask.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], <vscale x 8 x i32> [[VS1]], <vscale x 8 x i32> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmacc_vv_i64m8_m(vbool8_t mask, vint64m8_t vd, vint32m4_t vs1, vint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vv_i64m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmacc_vx_i64m8_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmacc.mask.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmacc_vx_i64m8_m(vbool8_t mask, vint64m8_t vd, int32_t rs1, vint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmacc_vx_i64m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccsu.c b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccsu.c
new file mode 100644
index 00000000000000..3a36e546ed4608
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccsu.c
@@ -0,0 +1,367 @@
+// RUN: %clang_cc1 -triple riscv64 -target-feature +xtheadvector \
+// RUN:   -disable-O0-optnone -emit-llvm %s -o - | \
+// RUN:   opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccsu_vv_i16m2
+// CHECK-RV64-SAME: (<vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x i8> [[VS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccsu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x i8> [[VS1]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmaccsu_vv_i16m2(vint16m2_t vd, vint8m1_t vs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccsu_vx_i16m2
+// CHECK-RV64-SAME: (<vscale x 8 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccsu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmaccsu_vx_i16m2(vint16m2_t vd, int8_t rs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccsu_vv_i16m4
+// CHECK-RV64-SAME: (<vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x i8> [[VS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccsu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x i8> [[VS1]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmaccsu_vv_i16m4(vint16m4_t vd, vint8m2_t vs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccsu_vx_i16m4
+// CHECK-RV64-SAME: (<vscale x 16 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccsu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmaccsu_vx_i16m4(vint16m4_t vd, int8_t rs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccsu_vv_i16m8
+// CHECK-RV64-SAME: (<vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x i8> [[VS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccsu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x i8> [[VS1]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmaccsu_vv_i16m8(vint16m8_t vd, vint8m4_t vs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccsu_vx_i16m8
+// CHECK-RV64-SAME: (<vscale x 32 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccsu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmaccsu_vx_i16m8(vint16m8_t vd, int8_t rs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccsu_vv_i32m2
+// CHECK-RV64-SAME: (<vscale x 4 x i32> [[VD:%.*]], <vscale x 4 x i16> [[VS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccsu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], <vscale x 4 x i16> [[VS1]], <vscale x 4 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmaccsu_vv_i32m2(vint32m2_t vd, vint16m1_t vs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccsu_vx_i32m2
+// CHECK-RV64-SAME: (<vscale x 4 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccsu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmaccsu_vx_i32m2(vint32m2_t vd, int16_t rs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i32m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccsu_vv_i32m4
+// CHECK-RV64-SAME: (<vscale x 8 x i32> [[VD:%.*]], <vscale x 8 x i16> [[VS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccsu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], <vscale x 8 x i16> [[VS1]], <vscale x 8 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmaccsu_vv_i32m4(vint32m4_t vd, vint16m2_t vs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccsu_vx_i32m4
+// CHECK-RV64-SAME: (<vscale x 8 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccsu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmaccsu_vx_i32m4(vint32m4_t vd, int16_t rs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i32m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccsu_vv_i32m8
+// CHECK-RV64-SAME: (<vscale x 16 x i32> [[VD:%.*]], <vscale x 16 x i16> [[VS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccsu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], <vscale x 16 x i16> [[VS1]], <vscale x 16 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmaccsu_vv_i32m8(vint32m8_t vd, vint16m4_t vs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccsu_vx_i32m8
+// CHECK-RV64-SAME: (<vscale x 16 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccsu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmaccsu_vx_i32m8(vint32m8_t vd, int16_t rs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i32m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccsu_vv_i64m2
+// CHECK-RV64-SAME: (<vscale x 2 x i64> [[VD:%.*]], <vscale x 2 x i32> [[VS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccsu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], <vscale x 2 x i32> [[VS1]], <vscale x 2 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmaccsu_vv_i64m2(vint64m2_t vd, vint32m1_t vs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i64m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccsu_vx_i64m2
+// CHECK-RV64-SAME: (<vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccsu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmaccsu_vx_i64m2(vint64m2_t vd, int32_t rs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i64m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccsu_vv_i64m4
+// CHECK-RV64-SAME: (<vscale x 4 x i64> [[VD:%.*]], <vscale x 4 x i32> [[VS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccsu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], <vscale x 4 x i32> [[VS1]], <vscale x 4 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmaccsu_vv_i64m4(vint64m4_t vd, vint32m2_t vs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i64m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccsu_vx_i64m4
+// CHECK-RV64-SAME: (<vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccsu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmaccsu_vx_i64m4(vint64m4_t vd, int32_t rs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i64m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccsu_vv_i64m8
+// CHECK-RV64-SAME: (<vscale x 8 x i64> [[VD:%.*]], <vscale x 8 x i32> [[VS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccsu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], <vscale x 8 x i32> [[VS1]], <vscale x 8 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmaccsu_vv_i64m8(vint64m8_t vd, vint32m4_t vs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i64m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccsu_vx_i64m8
+// CHECK-RV64-SAME: (<vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccsu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmaccsu_vx_i64m8(vint64m8_t vd, int32_t rs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i64m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccsu_vv_i16m2_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x i8> [[VS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccsu.mask.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x i8> [[VS1]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmaccsu_vv_i16m2_m(vbool8_t mask, vint16m2_t vd, vint8m1_t vs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccsu_vx_i16m2_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccsu.mask.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmaccsu_vx_i16m2_m(vbool8_t mask, vint16m2_t vd, int8_t rs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccsu_vv_i16m4_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x i8> [[VS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccsu.mask.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x i8> [[VS1]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmaccsu_vv_i16m4_m(vbool4_t mask, vint16m4_t vd, vint8m2_t vs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccsu_vx_i16m4_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccsu.mask.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmaccsu_vx_i16m4_m(vbool4_t mask, vint16m4_t vd, int8_t rs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccsu_vv_i16m8_m
+// CHECK-RV64-SAME: (<vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x i8> [[VS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccsu.mask.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x i8> [[VS1]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmaccsu_vv_i16m8_m(vbool2_t mask, vint16m8_t vd, vint8m4_t vs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccsu_vx_i16m8_m
+// CHECK-RV64-SAME: (<vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i16> [[VD:%.*]], i8 noundef signext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccsu.mask.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmaccsu_vx_i16m8_m(vbool2_t mask, vint16m8_t vd, int8_t rs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i16m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccsu_vv_i32m2_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i32> [[VD:%.*]], <vscale x 4 x i16> [[VS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccsu.mask.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], <vscale x 4 x i16> [[VS1]], <vscale x 4 x i16> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmaccsu_vv_i32m2_m(vbool16_t mask, vint32m2_t vd, vint16m1_t vs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i32m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccsu_vx_i32m2_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccsu.mask.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmaccsu_vx_i32m2_m(vbool16_t mask, vint32m2_t vd, int16_t rs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i32m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccsu_vv_i32m4_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i32> [[VD:%.*]], <vscale x 8 x i16> [[VS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccsu.mask.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], <vscale x 8 x i16> [[VS1]], <vscale x 8 x i16> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmaccsu_vv_i32m4_m(vbool8_t mask, vint32m4_t vd, vint16m2_t vs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i32m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccsu_vx_i32m4_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccsu.mask.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmaccsu_vx_i32m4_m(vbool8_t mask, vint32m4_t vd, int16_t rs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i32m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccsu_vv_i32m8_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i32> [[VD:%.*]], <vscale x 16 x i16> [[VS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccsu.mask.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], <vscale x 16 x i16> [[VS1]], <vscale x 16 x i16> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmaccsu_vv_i32m8_m(vbool4_t mask, vint32m8_t vd, vint16m4_t vs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i32m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccsu_vx_i32m8_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i32> [[VD:%.*]], i16 noundef signext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccsu.mask.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmaccsu_vx_i32m8_m(vbool4_t mask, vint32m8_t vd, int16_t rs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i32m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccsu_vv_i64m2_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i64> [[VD:%.*]], <vscale x 2 x i32> [[VS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccsu.mask.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], <vscale x 2 x i32> [[VS1]], <vscale x 2 x i32> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmaccsu_vv_i64m2_m(vbool32_t mask, vint64m2_t vd, vint32m1_t vs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i64m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccsu_vx_i64m2_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccsu.mask.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmaccsu_vx_i64m2_m(vbool32_t mask, vint64m2_t vd, int32_t rs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i64m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccsu_vv_i64m4_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i64> [[VD:%.*]], <vscale x 4 x i32> [[VS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccsu.mask.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], <vscale x 4 x i32> [[VS1]], <vscale x 4 x i32> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmaccsu_vv_i64m4_m(vbool16_t mask, vint64m4_t vd, vint32m2_t vs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i64m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccsu_vx_i64m4_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccsu.mask.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmaccsu_vx_i64m4_m(vbool16_t mask, vint64m4_t vd, int32_t rs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i64m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccsu_vv_i64m8_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i64> [[VD:%.*]], <vscale x 8 x i32> [[VS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccsu.mask.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], <vscale x 8 x i32> [[VS1]], <vscale x 8 x i32> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmaccsu_vv_i64m8_m(vbool8_t mask, vint64m8_t vd, vint32m4_t vs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vv_i64m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccsu_vx_i64m8_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccsu.mask.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmaccsu_vx_i64m8_m(vbool8_t mask, vint64m8_t vd, int32_t rs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccsu_vx_i64m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccu.c b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccu.c
new file mode 100644
index 00000000000000..e568981e877e5b
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccu.c
@@ -0,0 +1,367 @@
+// RUN: %clang_cc1 -triple riscv64 -target-feature +xtheadvector \
+// RUN:   -disable-O0-optnone -emit-llvm %s -o - | \
+// RUN:   opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccu_vv_u16m2
+// CHECK-RV64-SAME: (<vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x i8> [[VS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccu.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x i8> [[VS1]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vwmaccu_vv_u16m2(vuint16m2_t vd, vuint8m1_t vs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u16m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccu_vx_u16m2
+// CHECK-RV64-SAME: (<vscale x 8 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccu.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vwmaccu_vx_u16m2(vuint16m2_t vd, uint8_t rs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccu_vv_u16m4
+// CHECK-RV64-SAME: (<vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x i8> [[VS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccu.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x i8> [[VS1]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vwmaccu_vv_u16m4(vuint16m4_t vd, vuint8m2_t vs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u16m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccu_vx_u16m4
+// CHECK-RV64-SAME: (<vscale x 16 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccu.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vwmaccu_vx_u16m4(vuint16m4_t vd, uint8_t rs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccu_vv_u16m8
+// CHECK-RV64-SAME: (<vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x i8> [[VS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccu.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x i8> [[VS1]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vwmaccu_vv_u16m8(vuint16m8_t vd, vuint8m4_t vs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u16m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccu_vx_u16m8
+// CHECK-RV64-SAME: (<vscale x 32 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccu.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vwmaccu_vx_u16m8(vuint16m8_t vd, uint8_t rs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccu_vv_u32m2
+// CHECK-RV64-SAME: (<vscale x 4 x i32> [[VD:%.*]], <vscale x 4 x i16> [[VS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccu.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], <vscale x 4 x i16> [[VS1]], <vscale x 4 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vuint32m2_t test_vwmaccu_vv_u32m2(vuint32m2_t vd, vuint16m1_t vs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u32m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccu_vx_u32m2
+// CHECK-RV64-SAME: (<vscale x 4 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccu.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vuint32m2_t test_vwmaccu_vx_u32m2(vuint32m2_t vd, uint16_t rs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u32m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccu_vv_u32m4
+// CHECK-RV64-SAME: (<vscale x 8 x i32> [[VD:%.*]], <vscale x 8 x i16> [[VS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccu.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], <vscale x 8 x i16> [[VS1]], <vscale x 8 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vuint32m4_t test_vwmaccu_vv_u32m4(vuint32m4_t vd, vuint16m2_t vs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u32m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccu_vx_u32m4
+// CHECK-RV64-SAME: (<vscale x 8 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccu.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vuint32m4_t test_vwmaccu_vx_u32m4(vuint32m4_t vd, uint16_t rs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u32m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccu_vv_u32m8
+// CHECK-RV64-SAME: (<vscale x 16 x i32> [[VD:%.*]], <vscale x 16 x i16> [[VS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccu.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], <vscale x 16 x i16> [[VS1]], <vscale x 16 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vuint32m8_t test_vwmaccu_vv_u32m8(vuint32m8_t vd, vuint16m4_t vs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u32m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccu_vx_u32m8
+// CHECK-RV64-SAME: (<vscale x 16 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccu.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vuint32m8_t test_vwmaccu_vx_u32m8(vuint32m8_t vd, uint16_t rs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u32m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccu_vv_u64m2
+// CHECK-RV64-SAME: (<vscale x 2 x i64> [[VD:%.*]], <vscale x 2 x i32> [[VS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccu.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], <vscale x 2 x i32> [[VS1]], <vscale x 2 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vuint64m2_t test_vwmaccu_vv_u64m2(vuint64m2_t vd, vuint32m1_t vs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u64m2(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccu_vx_u64m2
+// CHECK-RV64-SAME: (<vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccu.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vuint64m2_t test_vwmaccu_vx_u64m2(vuint64m2_t vd, uint32_t rs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u64m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccu_vv_u64m4
+// CHECK-RV64-SAME: (<vscale x 4 x i64> [[VD:%.*]], <vscale x 4 x i32> [[VS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccu.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], <vscale x 4 x i32> [[VS1]], <vscale x 4 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vuint64m4_t test_vwmaccu_vv_u64m4(vuint64m4_t vd, vuint32m2_t vs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u64m4(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccu_vx_u64m4
+// CHECK-RV64-SAME: (<vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccu.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vuint64m4_t test_vwmaccu_vx_u64m4(vuint64m4_t vd, uint32_t rs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u64m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccu_vv_u64m8
+// CHECK-RV64-SAME: (<vscale x 8 x i64> [[VD:%.*]], <vscale x 8 x i32> [[VS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccu.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], <vscale x 8 x i32> [[VS1]], <vscale x 8 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vuint64m8_t test_vwmaccu_vv_u64m8(vuint64m8_t vd, vuint32m4_t vs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u64m8(vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccu_vx_u64m8
+// CHECK-RV64-SAME: (<vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccu.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vuint64m8_t test_vwmaccu_vx_u64m8(vuint64m8_t vd, uint32_t rs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u64m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccu_vv_u16m2_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i16> [[VD:%.*]], <vscale x 8 x i8> [[VS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccu.mask.nxv8i16.nxv8i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], <vscale x 8 x i8> [[VS1]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vwmaccu_vv_u16m2_m(vbool8_t mask, vuint16m2_t vd, vuint8m1_t vs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u16m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccu_vx_u16m2_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccu.mask.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vuint16m2_t test_vwmaccu_vx_u16m2_m(vbool8_t mask, vuint16m2_t vd, uint8_t rs1, vuint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccu_vv_u16m4_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i16> [[VD:%.*]], <vscale x 16 x i8> [[VS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccu.mask.nxv16i16.nxv16i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], <vscale x 16 x i8> [[VS1]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vwmaccu_vv_u16m4_m(vbool4_t mask, vuint16m4_t vd, vuint8m2_t vs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u16m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccu_vx_u16m4_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccu.mask.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vuint16m4_t test_vwmaccu_vx_u16m4_m(vbool4_t mask, vuint16m4_t vd, uint8_t rs1, vuint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccu_vv_u16m8_m
+// CHECK-RV64-SAME: (<vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i16> [[VD:%.*]], <vscale x 32 x i8> [[VS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccu.mask.nxv32i16.nxv32i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], <vscale x 32 x i8> [[VS1]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vwmaccu_vv_u16m8_m(vbool2_t mask, vuint16m8_t vd, vuint8m4_t vs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u16m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccu_vx_u16m8_m
+// CHECK-RV64-SAME: (<vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccu.mask.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vuint16m8_t test_vwmaccu_vx_u16m8_m(vbool2_t mask, vuint16m8_t vd, uint8_t rs1, vuint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u16m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccu_vv_u32m2_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i32> [[VD:%.*]], <vscale x 4 x i16> [[VS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccu.mask.nxv4i32.nxv4i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], <vscale x 4 x i16> [[VS1]], <vscale x 4 x i16> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vuint32m2_t test_vwmaccu_vv_u32m2_m(vbool16_t mask, vuint32m2_t vd, vuint16m1_t vs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u32m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccu_vx_u32m2_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccu.mask.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vuint32m2_t test_vwmaccu_vx_u32m2_m(vbool16_t mask, vuint32m2_t vd, uint16_t rs1, vuint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u32m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccu_vv_u32m4_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i32> [[VD:%.*]], <vscale x 8 x i16> [[VS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccu.mask.nxv8i32.nxv8i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], <vscale x 8 x i16> [[VS1]], <vscale x 8 x i16> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vuint32m4_t test_vwmaccu_vv_u32m4_m(vbool8_t mask, vuint32m4_t vd, vuint16m2_t vs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u32m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccu_vx_u32m4_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccu.mask.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vuint32m4_t test_vwmaccu_vx_u32m4_m(vbool8_t mask, vuint32m4_t vd, uint16_t rs1, vuint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u32m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccu_vv_u32m8_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i32> [[VD:%.*]], <vscale x 16 x i16> [[VS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccu.mask.nxv16i32.nxv16i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], <vscale x 16 x i16> [[VS1]], <vscale x 16 x i16> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vuint32m8_t test_vwmaccu_vv_u32m8_m(vbool4_t mask, vuint32m8_t vd, vuint16m4_t vs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u32m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccu_vx_u32m8_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccu.mask.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vuint32m8_t test_vwmaccu_vx_u32m8_m(vbool4_t mask, vuint32m8_t vd, uint16_t rs1, vuint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u32m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccu_vv_u64m2_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i64> [[VD:%.*]], <vscale x 2 x i32> [[VS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccu.mask.nxv2i64.nxv2i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], <vscale x 2 x i32> [[VS1]], <vscale x 2 x i32> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vuint64m2_t test_vwmaccu_vv_u64m2_m(vbool32_t mask, vuint64m2_t vd, vuint32m1_t vs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u64m2_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccu_vx_u64m2_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccu.mask.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vuint64m2_t test_vwmaccu_vx_u64m2_m(vbool32_t mask, vuint64m2_t vd, uint32_t rs1, vuint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u64m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccu_vv_u64m4_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i64> [[VD:%.*]], <vscale x 4 x i32> [[VS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccu.mask.nxv4i64.nxv4i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], <vscale x 4 x i32> [[VS1]], <vscale x 4 x i32> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vuint64m4_t test_vwmaccu_vv_u64m4_m(vbool16_t mask, vuint64m4_t vd, vuint32m2_t vs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u64m4_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccu_vx_u64m4_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccu.mask.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vuint64m4_t test_vwmaccu_vx_u64m4_m(vbool16_t mask, vuint64m4_t vd, uint32_t rs1, vuint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u64m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccu_vv_u64m8_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i64> [[VD:%.*]], <vscale x 8 x i32> [[VS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccu.mask.nxv8i64.nxv8i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], <vscale x 8 x i32> [[VS1]], <vscale x 8 x i32> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vuint64m8_t test_vwmaccu_vv_u64m8_m(vbool8_t mask, vuint64m8_t vd, vuint32m4_t vs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vv_u64m8_m(mask, vd, vs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccu_vx_u64m8_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccu.mask.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vuint64m8_t test_vwmaccu_vx_u64m8_m(vbool8_t mask, vuint64m8_t vd, uint32_t rs1, vuint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccu_vx_u64m8_m(mask, vd, rs1, vs2, vl);
+}
+
diff --git a/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccus.c b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccus.c
new file mode 100644
index 00000000000000..4d53e7d09201cb
--- /dev/null
+++ b/clang/test/CodeGen/RISCV/rvv0p71-intrinsics-handcrafted/vector-widening-mul-add/thead/vwmaccus.c
@@ -0,0 +1,187 @@
+// RUN: %clang_cc1 -triple riscv64 -target-feature +xtheadvector \
+// RUN:   -disable-O0-optnone -emit-llvm %s -o - | \
+// RUN:   opt -S -passes=mem2reg | \
+// RUN:   FileCheck --check-prefix=CHECK-RV64 %s
+
+#include <riscv_vector.h>
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccus_vx_i16m2
+// CHECK-RV64-SAME: (<vscale x 8 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccus.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmaccus_vx_i16m2(vint16m2_t vd, uint8_t rs1, vint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i16m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccus_vx_i16m4
+// CHECK-RV64-SAME: (<vscale x 16 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccus.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmaccus_vx_i16m4(vint16m4_t vd, uint8_t rs1, vint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i16m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccus_vx_i16m8
+// CHECK-RV64-SAME: (<vscale x 32 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccus.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmaccus_vx_i16m8(vint16m8_t vd, uint8_t rs1, vint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i16m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccus_vx_i32m2
+// CHECK-RV64-SAME: (<vscale x 4 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccus.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmaccus_vx_i32m2(vint32m2_t vd, uint16_t rs1, vint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i32m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccus_vx_i32m4
+// CHECK-RV64-SAME: (<vscale x 8 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccus.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmaccus_vx_i32m4(vint32m4_t vd, uint16_t rs1, vint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i32m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccus_vx_i32m8
+// CHECK-RV64-SAME: (<vscale x 16 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccus.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmaccus_vx_i32m8(vint32m8_t vd, uint16_t rs1, vint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i32m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccus_vx_i64m2
+// CHECK-RV64-SAME: (<vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccus.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmaccus_vx_i64m2(vint64m2_t vd, uint32_t rs1, vint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i64m2(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccus_vx_i64m4
+// CHECK-RV64-SAME: (<vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccus.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmaccus_vx_i64m4(vint64m4_t vd, uint32_t rs1, vint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i64m4(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccus_vx_i64m8
+// CHECK-RV64-SAME: (<vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccus.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmaccus_vx_i64m8(vint64m8_t vd, uint32_t rs1, vint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i64m8(vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i16> @test_vwmaccus_vx_i16m2_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 8 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i16> @llvm.riscv.th.vwmaccus.mask.nxv8i16.i8.nxv8i8.i64(<vscale x 8 x i16> [[VD]], i8 [[RS1]], <vscale x 8 x i8> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i16> [[TMP0]]
+//
+vint16m2_t test_vwmaccus_vx_i16m2_m(vbool8_t mask, vint16m2_t vd, uint8_t rs1, vint8m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i16m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i16> @test_vwmaccus_vx_i16m4_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 16 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i16> @llvm.riscv.th.vwmaccus.mask.nxv16i16.i8.nxv16i8.i64(<vscale x 16 x i16> [[VD]], i8 [[RS1]], <vscale x 16 x i8> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i16> [[TMP0]]
+//
+vint16m4_t test_vwmaccus_vx_i16m4_m(vbool4_t mask, vint16m4_t vd, uint8_t rs1, vint8m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i16m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 32 x i16> @test_vwmaccus_vx_i16m8_m
+// CHECK-RV64-SAME: (<vscale x 32 x i1> [[MASK:%.*]], <vscale x 32 x i16> [[VD:%.*]], i8 noundef zeroext [[RS1:%.*]], <vscale x 32 x i8> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 32 x i16> @llvm.riscv.th.vwmaccus.mask.nxv32i16.i8.nxv32i8.i64(<vscale x 32 x i16> [[VD]], i8 [[RS1]], <vscale x 32 x i8> [[VS2]], <vscale x 32 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 32 x i16> [[TMP0]]
+//
+vint16m8_t test_vwmaccus_vx_i16m8_m(vbool2_t mask, vint16m8_t vd, uint8_t rs1, vint8m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i16m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i32> @test_vwmaccus_vx_i32m2_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 4 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i32> @llvm.riscv.th.vwmaccus.mask.nxv4i32.i16.nxv4i16.i64(<vscale x 4 x i32> [[VD]], i16 [[RS1]], <vscale x 4 x i16> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i32> [[TMP0]]
+//
+vint32m2_t test_vwmaccus_vx_i32m2_m(vbool16_t mask, vint32m2_t vd, uint16_t rs1, vint16m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i32m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i32> @test_vwmaccus_vx_i32m4_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 8 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i32> @llvm.riscv.th.vwmaccus.mask.nxv8i32.i16.nxv8i16.i64(<vscale x 8 x i32> [[VD]], i16 [[RS1]], <vscale x 8 x i16> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i32> [[TMP0]]
+//
+vint32m4_t test_vwmaccus_vx_i32m4_m(vbool8_t mask, vint32m4_t vd, uint16_t rs1, vint16m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i32m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 16 x i32> @test_vwmaccus_vx_i32m8_m
+// CHECK-RV64-SAME: (<vscale x 16 x i1> [[MASK:%.*]], <vscale x 16 x i32> [[VD:%.*]], i16 noundef zeroext [[RS1:%.*]], <vscale x 16 x i16> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 16 x i32> @llvm.riscv.th.vwmaccus.mask.nxv16i32.i16.nxv16i16.i64(<vscale x 16 x i32> [[VD]], i16 [[RS1]], <vscale x 16 x i16> [[VS2]], <vscale x 16 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 16 x i32> [[TMP0]]
+//
+vint32m8_t test_vwmaccus_vx_i32m8_m(vbool4_t mask, vint32m8_t vd, uint16_t rs1, vint16m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i32m8_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 2 x i64> @test_vwmaccus_vx_i64m2_m
+// CHECK-RV64-SAME: (<vscale x 2 x i1> [[MASK:%.*]], <vscale x 2 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 2 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 2 x i64> @llvm.riscv.th.vwmaccus.mask.nxv2i64.i32.nxv2i32.i64(<vscale x 2 x i64> [[VD]], i32 [[RS1]], <vscale x 2 x i32> [[VS2]], <vscale x 2 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 2 x i64> [[TMP0]]
+//
+vint64m2_t test_vwmaccus_vx_i64m2_m(vbool32_t mask, vint64m2_t vd, uint32_t rs1, vint32m1_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i64m2_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 4 x i64> @test_vwmaccus_vx_i64m4_m
+// CHECK-RV64-SAME: (<vscale x 4 x i1> [[MASK:%.*]], <vscale x 4 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 4 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 4 x i64> @llvm.riscv.th.vwmaccus.mask.nxv4i64.i32.nxv4i32.i64(<vscale x 4 x i64> [[VD]], i32 [[RS1]], <vscale x 4 x i32> [[VS2]], <vscale x 4 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 4 x i64> [[TMP0]]
+//
+vint64m4_t test_vwmaccus_vx_i64m4_m(vbool16_t mask, vint64m4_t vd, uint32_t rs1, vint32m2_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i64m4_m(mask, vd, rs1, vs2, vl);
+}
+
+// CHECK-RV64-LABEL: define dso_local <vscale x 8 x i64> @test_vwmaccus_vx_i64m8_m
+// CHECK-RV64-SAME: (<vscale x 8 x i1> [[MASK:%.*]], <vscale x 8 x i64> [[VD:%.*]], i32 noundef signext [[RS1:%.*]], <vscale x 8 x i32> [[VS2:%.*]], i64 noundef [[VL:%.*]]) #[[ATTR0:[0-9]+]] {
+// CHECK-RV64-NEXT:  entry:
+// CHECK-RV64-NEXT:    [[TMP0:%.*]] = call <vscale x 8 x i64> @llvm.riscv.th.vwmaccus.mask.nxv8i64.i32.nxv8i32.i64(<vscale x 8 x i64> [[VD]], i32 [[RS1]], <vscale x 8 x i32> [[VS2]], <vscale x 8 x i1> [[MASK]], i64 [[VL]])
+// CHECK-RV64-NEXT:    ret <vscale x 8 x i64> [[TMP0]]
+//
+vint64m8_t test_vwmaccus_vx_i64m8_m(vbool8_t mask, vint64m8_t vd, uint32_t rs1, vint32m4_t vs2, size_t vl) {
+  return __riscv_th_vwmaccus_vx_i64m8_m(mask, vd, rs1, vs2, vl);
+}
+